 |
|
The Genetics of Psoriasis 2001
The Odyssey Continues
James T. Elder, MD, PhD;
Rajan P. Nair, PhD;
Tilo Henseler, MD, PhD;
Stefan Jenisch, MD;
Philip Stuart, BS;
Nicholas Chia, MD;
Enno Christophers, MD;
John J. Voorhees, MD
Arch Dermatol. 2001;137:1447-1454.
ABSTRACT
| |
Accumulating evidence indicates that psoriasis is a multifactorial disorder caused by the concerted action of multiple disease genes in a single individual, triggered by environmental factors. Some of these genes control the severity of multiple diseases by regulating inflammation and immunity (severity genes), whereas others are unique to psoriasis. Various combinations of these genes can occur even within a single family, accounting in large measure for the many clinical manifestations of psoriasis. The disease-causing variants (alleles) of these genes probably arose early in the history of modern humans. As a result, psoriasis disease alleles are common in the general population, have a worldwide distribution, and often share the same ancestral chromosome with neutral alleles at adjacent loci. This phenomenon, called linkage disequilibrium, explains why psoriasis is strongly associated with HLA-Cw6 worldwide, although HLA-Cw6 is unlikely to be the disease allele. Many unaffected individuals carry 1 or more disease alleles, but lack other genetic and/or environmental factors necessary to produce disease. This explains why psoriasis develops in only about 10% of HLA-Cw6–positive individuals, and why genome-wide linkage scans for psoriasis and other multifactorial genetic disorders have not been uniformly successful. The Human Genome Project is rapidly generating a catalog of human DNA sequence variations. This resource has already allowed precise linkage disequilibrium mapping of the major histocompatibility complex psoriasis gene to just beyond HLA-C, toward HLA-A. This gene is likely to be identified soon. Further development and use of linkage disequilibrium resources will provide a powerful tool for the identification of the remaining psoriasis genes.
INTRODUCTION
Psoriasis is a chronic, inflammatory, hyperproliferative disease of the skin, scalp, nails, and joints, affecting 1% to 2% of the US population at an estimated cost of more than $3 billion per year.1 It has a variety of clinical presentations, most of which eventuate into erythematous, scaly plaques with or without nail disease and arthritis. At the cellular level, psoriasis is characterized by (1) markedly increased epidermal proliferation and incomplete differentiation; (2) elongation, dilatation, and "leakiness" of the superficial plexus of dermal capillaries; and (3) a mixed inflammatory and immune cell infiltrate of the epidermis and papillary dermis. Psoriasis is found worldwide, although its frequency varies widely among different ethnic groups. Susceptibility to psoriasis is unmistakably heritable, but environmental factors, notably trauma, stress, and infections, are also important determinants of disease onset and severity. While several plausible immune/inflammatory mechanisms have been proposed,2 true molecular insight into the cause of psoriasis is lacking.
This review will address the following 4 points of relevance to clinicians and basic scientists alike: (1) How can we be sure psoriasis is genetic? (2) What do we know about PSORS1 (psoriasis susceptibility 1), the psoriasis gene in the major histocompatibility complex (MHC)? (3) What do we know about psoriasis genes across the rest of the genome? and (4) Do genetic differences explain the many clinical variants of psoriasis?
HOW CAN WE BE SURE PSORIASIS IS GENETIC?
Psoriasis only appears to run in families about a third of the time. Even in those families, it often does not appear to follow simple patterns of dominant or recessive inheritance.3-5 The simplest way to explain this observation is to assume that disease alleles (ie, alternative forms of a gene at a given locus) encoded by several genes located at several different loci throughout the genome are required for the disease to develop. The model depicted in Figure 1 suggests that disease alleles at each of 4 loci on different chromosomes are required for psoriasis to develop, and that only 1 copy of each disease allele is required for the disease to develop (ie, each disease allele acts in a dominant fashion). Although this model is unproven, it is instructive in explaining the low frequency of families with multiple psoriatic members. Under this model, the chance of psoriasis developing in any given individual would be ( )4 =  . Thus, families would need to be quite large to observe more than 1 case per family. The fact that familial involvement is observed in one third of cases suggests that the burden of psoriasis genes in the healthy population may be even higher than that shown in Figure 1.
|
|
|
|
Figure 1. Polygenic model of psoriasis. Four psoriasis genes (PS1-PS4) are depicted on 4 different chromosomes, contributing to psoriasis (solid symbol) in a simple pedigree. In the model shown, each allele acts in a dominant fashion to produce disease, but only when disease alleles at the other 3 loci are present. Circles indicate female members; squares, male members.
|
|
|
What is the actual evidence of the genetic basis for psoriasis? One of the best indications we have comes from twin studies. Identical twins share all their genes in common, whereas fraternal twins, like siblings, only share half their genes in common. Thus, even if a disorder is caused by many genes, disease should be concordant (present in both) in identical twins more often than in fraternal twins. Psoriasis fulfills this requirement, being concordant about 3 times as often in identical twins as in fraternal twins. Both concordance rates, however, are markedly lower in Australia than in the United States or in Denmark.6-8 As solar radiation is more intense in Australia than in Denmark and most of the United States, these differences suggest that ambient UV irradiation may be a therapeutic environmental factor in Australia. Twin studies and pedigree studies can be used to calculate heritability (h2), a value used to measure the proportion of variability of a trait that is due to genetic factors. As reviewed previously,3 various estimates have placed the heritability of psoriasis in the range of 60% to 90%, among the highest of the multifactorial genetic disorders (ie, diseases caused by more than 1 gene plus environmental factors).
Another important measure of the genetic character of psoriasis is the risk ratio. For any set of probands (ie, index case in a family), the risk ratio is defined as the prevalence of a disease in relatives of a given degree of relatedness divided by the prevalence of disease in the general population. (First-degree relatives share half of their genetic material; second-degree relatives, one fourth; etc.) On the basis of a large epidemiological study, Henseler and Christophers9 showed that the risk ratio is approximately 10 for first-degree relatives of patients with juvenile-onset psoriasis (Figure 2). This finding is very similar to those of the classic epidemiological studies of Lomholt11 and Hellgren,12 as reviewed previously.3 This finding has major significance for studies of genetic linkage of psoriasis (ie, the measure of the correlation between transmission of any allele at the marker locus and the disease locus within families). It has been shown that if the risk ratio for first-degree relatives is at least 4, and at least 1 of the loci is of major effect, then a search for the genes by means of genetic linkage techniques is feasible.13 These conditions appear to be met in the case of juvenile-onset psoriasis, if one assumes a strong genetic effect coming from within the MHC. Therefore, a decade ago, we began searching for the psoriasis genes by means of genetic linkage in families, requiring that the proband must have juvenile-onset psoriasis. Our studies indicate that more than 90% of psoriatic relatives of the proband also have juvenile-onset disease. In contrast, only about 75% of psoriatic patients in general fall into the juvenile-onset category.9-10 This finding provides additional support for the genetic basis of juvenile-onset psoriasis.
|
|
|
|
Figure 2. Increased risk for psoriasis in relatives of psoriatic probands. The risk ratio is defined as the prevalence of psoriasis in the indicated type of relative, divided by the prevalence of psoriasis in the general population (assumed here to be 1.5%). Type I psoriasis has an age of onset of younger than 40 years; type II psoriasis, 40 years or older. Data are from Christophers and Henseler,10 modified from Elder et al.3
|
|
|
Because of the long-standing and strong evidence of HLA associations in psoriasis, our working hypothesis has been that 1 of the psoriasis genes resides in the MHC, on the short arm of chromosome 6, and that several other psoriasis genes are scattered throughout the human genome. Some of these genes may be general regulators of inflammation and/or immunity, and therefore may be involved in a variety of inflammatory and/or immune diseases in addition to psoriasis.14 These are termed severity genes. Other genes may make a distinctive contribution to the psoriatic phenotype; these are termed disease-specific genes.
WHAT DO WE KNOW ABOUT PSORS1, THE MHC PSORIASIS GENE?
Associations between psoriasis and particular HLA types have been recognized for nearly 30 years.15 Henseler and Christophers9 defined type I psoriasis as having age of onset younger than 40 years, with strong HLA associations. Patients with type II disease were characterized by age of onset 40 years or older, and much weaker HLA associations.9-10 Patients with type I disease showed a much stronger tendency for familial involvement. As shown in Figure 2, the risk ratio for first-degree relatives was approximately 10 for patients with type I psoriasis, but only about 1 to 2 for those with type II psoriasis.
The HLA associations identified in our study of familial psoriasis (Table 1) are very similar to those identified in previous case-control studies of HLA association in psoriasis.3 In particular, strong associations with HLA-Cw6 and HLA-B57 were found. We know that approximately two thirds of the psoriatic patients in those earlier studies had a family history negative for psoriasis and could be assumed to represent sporadic cases. The similarity of the HLA associations obtained in pedigree and case-control studies implies that so-called sporadic psoriasis must also have a genetic basis. Because no other inflammatory disease manifests such a strong HLA-C association, we suspect that the MHC psoriasis gene is a disease-specific gene.
|
|
|
|
Table 1. Selected HLA Associations With Familial Psoriasis in the Michigan-Kiel Study
|
|
|
In northern Europe and the United States, only about two thirds of psoriatic patients carry HLA-Cw6. This figure is even lower in most Oriental populations, where many affected individuals carry HLA-Cw7. In an effort to explain this observation, others have postulated a primary role for alanine at position 73 of the HLA-C protein molecule, which is also found on HLA-Cw7. We have not found such a role for alanine-73 in our study, as shown on the last line of Table 1. We believe that this phenomenon is better explained by the phenomenon of linkage disequilibrium.
Usually, HLA associations are due to linkage disequilibrium between the disease locus and 1 or more HLA loci. Linkage disequilibrium is a measure of the presence of a particular marker allele in apparently unrelated affected individuals. It usually represents a special case of genetic linkage, caused by the founder effect (ie, the occurrence of a disease-predisposing mutation in a distant ancestor). In the case of genetic linkage, different marker alleles are observed to segregate with disease from family to family, usually due to the occurrence of different mutations in the gene in question from family to family (Figure 3A). In contrast, when linkage disequilibrium is present, the same marker allele will segregate through seemingly unrelated pedigrees (Figure 3C). This occurs because the disease allele and the marker allele are descended from a single founder, and therefore still reside on the same ancestral chromosome. This phenomenon allows linkage disequilibrium (but not linkage) to be detected in a case-control sample (Figure 3B). As generations pass, the segment of the ancestral chromosome containing the disease mutation becomes shorter and shorter, due to meiotic recombination events occurring in multiple generations. In contrast, genetic linkage is only broken by recombination events occurring within an individual family, which in practice rarely contains more than 3 generations and therefore presents far less opportunity for meiotic recombination. Thus, linkage disequilibrium is a powerful tool for fine mapping of genes, if one can first find the approximate location of the gene by means of genetic linkage.
|
|
|
|
Figure 3. Examples of linkage (A), association (B), and association and linkage disequilibrium in families (C). The chromosome-shaped symbols depict a marker locus on top, and the disease gene below. The marker allele assumes one of several values, and the disease gene can encode a disease allele (PS) or a normal allele (N).
|
|
|
Since 1997, several groups have presented evidence of genetic linkage between psoriasis and the MHC.17-19 However, despite strong evidence of HLA association (Table 1), it initially proved difficult for us to detect convincing, significant linkage to the HLA region in our original genome scan (ie, the technique whereby a collection of polymorphic markers distributed evenly across the genome are tested for genetic linkage with a particular trait).20 The MHC region (chromosome 6p21.3) contained one of the most significant linkage signals, but this peak did not reach the criterion lod scoreof 3.3 required for definitive declaration of linkage.21 (The term lod score refers to the logartihm of the odds ratio, a measure of genetic linkage, where the odds ratio is defined as the likelihood of encountering the observed outcome if the marker and disease are linked, divided by the likelihood of the same outcome if the marker and disease are unlinked.) Moreover, this peak was only observed under a recessive genetic model (Figure 4). This was surprising, because most psoriatic patients who carry HLA-Cw6 carry only 1 copy of this allele, and inheritance of this locus should therefore be dominant. The explanation for this difficulty in detecting linkage appears to be the presence of numerous unaffected married-in HLA-Cw6 carriers in our pedigrees, creating a situation in which the HLA-linked susceptibility gene is brought in from both sides of the family (Figure 5). When an affected member of the third generation inherits the susceptibility chromosome from an unaffected married-in family member (marked by ovals in Figure 5), the linkage computer program is forced to assume that a recombination event has occurred. As the computer detects more such apparent recombination events, the statistical evidence for linkage diminishes. In subsequent studies, we took this phenomenon into account using the presence of HLA-Cw6 to identify these unaffected carriers. Under these conditions, we were readily able to detect linkage to HLA-C, with a lod score of nearly 10.22
|
|
|
|
Figure 4. Results from our genome-wide scan for psoriasis genes under dominant (A) and recessive (B) genetic models. Linkage results expressed as lod scores are presented as a function of genetic distance along the chromosomes, which are aligned end-to-end with the short arm of each chromosome on the left. The peak corresponding to PSORS1 in the major histocompatibility complex (MHC) is indicated, along with several other linkage peaks that have also been observed in psoriasis and other inflammatory/immune disorders with skin involvement. cM indicates centimorgan. Modified from Nair et al20 by permission of Oxford University Press.
|
|
|
|
|
|
|
Figure 5. Transmission of HLA haplotypes in a familial psoriasis pedigree. Ovals demonstrate transmission of 2 different HLA-Cw6–bearing risk haplotypes from unaffected married-in pedigree members in the second generation. Rectangles demonstrate transmission of a third HLA-Cw6–bearing haplotype from an unaffected individual in the top generation. The transmission events marked with asterisks would be scored as recombinant by typical genetic linkage programs, thereby decreasing evidence of linkage. Circles indicate female members; squares, male members; solid shapes, affected members; slash, deceased; and shaded boxes, a recombination event between HLA-B and HLA-DR. Modified from Nair et al23 by permission of The University of Chicago Press.
|
|
|
The fact that we were able to establish linkage to HLA-C does not mean that HLA-C is itself PSORS1. The HLA region is nearly 3.5 million base pairs long and contains more than 200 genes. Because not all psoriatic patients carry HLA-Cw6, and because not all HLA-Cw6–bearing haplotypes (ie, the set of alleles carried by genes on 1 chromosome of a particular individual) confer equal risk for psoriasis, we suspected that the strong HLA-Cw6 association reflects a causative role for a nearby gene on the same ancestral chromosome. To address this possibility, we developed a dense microsatellite marker map of the HLA region and used it to more precisely map the psoriasis susceptibility determinant within the HLA region.23 (Microsatellites are short DNA sequences, usually repeats of C and A, that vary in length from chromosome to chromosome, and can therefore be used to distinguish between maternal and paternal chromosomes). The critical tool for localizing PSORS1 turned out to be ancestral haplotype analysis. In this method, haplotypes are inferred by following the segregation of markers through individual pedigrees, then these haplotypes are clustered to identify ancestral haplotypes shared between families. Numerous ancestral haplotypes were identified and tested for linkage disequilibrium with psoriasis. The shortest haplotypes that consistently demonstrated linkage disequilibrium shared a 60- to 70-kilobase (kb) interval extending from 30 to approximately 100 kb from HLA-C, toward the telomere (end) of the short arm of chromosome 6.23 The location of this haplotype was in good agreement with the regions of peak linkage disequilibrium identified by 2 other recent studies.24-25 We have named this 60-kb interval risk haplotype 1 (RH1) (Figure 6). Risk haplotype 1 did not contain HLA-Cw6 as a component allele, providing the strongest evidence to date that HLA-Cw6 is not PSORS1. The RH1 region is rich in repetitive sequences, including remnants of 3 different endogenous retroviruses. However, no genes have been identified in RH1 to date. There are several additional genes mapping just telomeric to RH1, including OTF3 (also known as POU5F1), TCF19 (also known as SC1), HCR (also known as PG8), and corneodesmosin (also known as the S gene) (Figure 6). Each of these genes displays strong associations with psoriasis.26-29 However, none of these genes predicts the risk for psoriasis as well as does the presence of RH1.27, 30-35 Therefore, we are continuing to focus on the RH1 region, although it appears to contain no genes. We are determining the DNA sequence across the RH1 region in a set of normal and disease chromosomes. By comparing these sequences, we hope to tease out the critical DNA sequence difference that predisposes to psoriasis.
|
|
|
|
Figure 6. Map of the major histocompatibility complex (MHC) class I region, showing the location of risk haplotype 1 (RH1) and nearby genes (solid bars) and pseudogenes (striped bars). Pseudogenes contain DNA sequence variations that prevent them from being produced as functional proteins. Distances from HLA-C have been determined from the DNA sequence of the MHC.
|
|
|
The MHC psoriasis gene, then, appears not to encode any of the known HLA antigens. Psoriasis therefore differs from several other MHC-linked disorders, including ankylosing spondylitis (AS),36 Behçet disease,37 and type 1 diabetes mellitus,38 in which the best available evidence indicates that the most closely disease-associated HLA alleles are directly responsible for disease susceptibility. The proof of this point awaits the discovery of the disease-predisposing mutation.
WHAT DO WE KNOW ABOUT THE NON-MHC PSORIASIS GENES?
Our studies have shown that RH1 is found in approximately 15% of the general population, and yet the prevalence of psoriasis is only approximately one tenth of this (1.5%). Other studies have repeatedly shown that only about 10% of those carrying HLA-Cw6 ever develop psoriasis.3 Why is this? Although environmental factors such as streptococcal infection and stress undoubtedly play a role, we believe that this is primarily because of a requirement for additional disease alleles, encoded by different genes in the same person, as envisioned in the model shown in Figure 1. Presumably, these genetic and environmental factors are not present in the right combination up to 90% of the time.
There have been numerous reports of non-MHC loci in psoriasis (Table 2). However, in contrast to the prevailing agreement on the importance of the PSORS1 locus, these additional non-MHC loci have only sometimes been confirmed by other groups, often without overwhelmingly strong evidence.20, 39, 42 In the field of complex disease genetics, a linkage result found by one group may not be confirmed by others, even if the linkage is real.47 In addition to PSORS1, at least 2 independent groups have found evidence of linkage of psoriasis to chromosomes 1q, 17q, and 20p and the central area of chromosome 19.17, 20, 40, 45, 48 These confirmatory findings are not definitive proof of the existence of a gene, but they are nevertheless very encouraging that non-MHC psoriasis genes do exist. We are currently collaborating with 2 other groups looking for psoriasis genes in an effort to learn whether several loci identified by various genome scans represent true positive results.
|
|
|
|
Table 2. Candidate Loci in Psoriasis Identified by Genetic Linkage Studies
|
|
|
We believe that the high frequency of disease alleles is one of the main reasons why the non-MHC loci have been difficult to identify and to replicate. If we assume that there are 3 non-MHC disease genes, all of equal frequency, and that these genes interact with PSORS1 in a multiplicative fashion (as might be the case if the products of the various genes acted along a common biochemical pathway), then each one of these non-MHC disease alleles would have to be present in approximately half of the general population! This scenario is depicted in Figure 7. There are now several multifactorial disorders for which strong evidence implicates a particular gene (or nearby markers in linkage disequilibrium with it). In many of these examples, the prevalence of the disease allele appears to be high. In androgenetic alopecia, a polymorphism in the androgen receptor gene is found in nearly 100% of young bald men, but also in 77% of old nonbald men.49 In type 1 diabetes mellitus, nearly 80% of the population carries a particular allele of the interleukin 12 p40 subunit gene that has been strongly implicated as a non-MHC locus in this disease.50 In 2 other examples (the calpain-10 gene in type 2 diabetes mellitus51 and the interleukin 3 gene in rheumatoid arthritis52), the frequency of the disease allele ranges from 20% to 50%. Thus, high disease allele frequencies appear to be the rule, rather than the exception, in the complex genetic disorders.
|
|
|
|
Figure 7. Hypothetical interaction for PSORS1 (psoriasis suseptibility 1) and the non–major histocompatibility complex genes in psoriasis, assuming a multiplicative model of gene-gene interaction consistent with the model of inheritance shown in Figure 1. The allele frequency of PSORS1 has been estimated on the basis of the frequency of HLA-Cw6 in white populations. The disease allele frequencies for genes X, Y, and Z are assumed to be equal. Note that high disease allele frequencies of genes X, Y, and Z are required to produce a final population prevalence of 2%.
|
|
|
Figure 8 is not an exhaustive depiction of all possible consequences of high gene frequency, but it does illustrate how high disease allele frequencies can make linkage harder to demonstrate. Part A depicts a dominant disease allele D being transferred from an affected father to 2 affected siblings. Linkage programs have no problem dealing with this situation. They can also be programmed to handle unaffected carriers (the father in part B) by including a variable telling the computer how often to expect unaffected carriers. However, they are stymied when 1 parent carries 2 copies of the disease allele (part C), or when both parents carry 1 or more copies (part D). In the situation depicted in part C, the computer cannot distinguish between disease alleles D and D', and therefore cannot determine whether either of these alleles is consistently passed along to the affected children. In part D, the affected child inherits the disease allele D' from the unaffected parent, rather than D from the affected parent. Unless the computer is told that the mother in part D is an unaffected carrier, it must (erroneously) conclude that D is not a relevant candidate for the disease gene. It is possible that PSORS1 might be the most readily detectable psoriasis gene by linkage techniques because it has a relatively low population prevalence (approximately 15%), thereby minimizing situations such as those depicted in parts C and D of Figure 8.
|
|
|
|
Figure 8. Deleterious effects of high disease allele frequencies on the ability to detect linkage. The examples shown are to be illustrative and not to depict all situations that can be encountered. D and D' indicate 2 genetically different disease alleles at a particular locus;N, any normal allele at that locus. Circles indicate female members; squares, male members; and solid shapes, affected members.
|
|
|
Given these difficulties, one might worry that the non-MHC genes might never be detected by tests of genetic linkage. However, the same feature of these genes that makes them hard to detect by tests of linkage (their high prevalence) may make it possible for us to detect them by tests of linkage disequilibrium. We have seen that if multiple genes contribute to disease in any given individual (as shown in Figure 1), then each disease allele must be quite prevalent in the population (Figure 6). On reflection, it becomes clear that the most plausible way that a disease allele can become common is for it to arise as a founder mutation relatively early in the history of modern humans, expanding thereafter as the human population expanded. In fact, many of these so-called disease alleles may actually be beneficial in other contexts, eg, a diabetes gene. One can readily imagine a scenario early in the history of modern humans in which food was scarce. In this setting, it was a good idea to hold on to and store as much glucose as possible, whenever a meal was available. However, in a world of plenty, possession of the same allele could result in diabetes. A founding father of human genetics, the late James Neel, MD, called this the thrifty gene hypothesis.53
As a result of such a founder effect, the same disease allele will be present in apparently unrelated families and individuals. If most of the complex genetic disorders are in fact caused by such founder mutations, then once a likely candidate gene is identified, it will be possible to test for them. Although several different tests of linkage disequilibrium are available, the most straightforward is the case-control study: one simply asks whether a specific gene variant is present more often in cases than in controls. Now that we have a fairly complete listing of some 30 000 genes in hand, courtesy of the Human Genome Project, the task of testing for candidate gene association is somewhat less daunting than it once seemed. However, other genes in the immediate vicinity of the candidate gene are also likely to exhibit strong disease associations, as we have observed for PSORS1 in the MHC. Therefore, some form of functional test will almost certainly be necessary to confirm the biological significance of each genetically detected candidate gene for psoriasis.
DO GENETIC DIFFERENCES EXPLAIN THE MANY CLINICAL VARIANTS OF PSORIASIS?
As dermatologists, we appreciate that guttate psoriasis is usually observed in younger individuals, and may or may not eventuate into chronic plaque psoriasis. From the work of Henseler and Christophers,9 we know that patients in whom psoriasis develops at a young age are more likely to carry HLA-Cw6, (and therefore to carry PSORS1). A recent study from England found that 100% of 29 patients with poststreptococcal guttate psoriasis were positive for HLA-Cw6.54 This finding suggests that the PSORS1 gene plays a major role in this form of psoriasis. This finding fits with the predominance of PSORS1 in juvenile-onset psoriasis, together with the fact that poststreptococcal guttate psoriasis is typically observed in younger individuals. In many of these individuals, disease resolves completely, without the evolution of chronic plaque disease. We would speculate that the presence or absence of additional genes may determine the patients in whom chronic plaque psoriasis develops. We have observed anecdotally that the few subjects who carried 2 copies of HLA-Cw6 tended to have severe and recalcitrant psoriasis, suggestive of a gene dosage effect. Similar gene dosage effects may be present for the non-MHC genes, singly or in various combinations.
The spondyloarthropathies can present with psoriasiform skin eruptions, suggesting certain common genetic determinants between these diseases. However, the joint manifestations of both are substantially different. HLA-Cw6 is detected substantially more often than HLA-B27 in psoriatic arthritis,55 indicating that HLA-B27 is not the major genetic determinant of psoriatic arthritis. As arthritis is a relatively infrequent concomitant of psoriasis, whereas it is a cardinal feature of AS, we would speculate that the primary genes determining psoriatic arthritis lie outside of the HLA region. The chromosome 17q locus is a prime candidate for such a gene, as 2 groups have found that the evidence of linkage to 17q is stronger in families that have joint symptoms or deformities.39, 42 The same region of chromosome 17q has also been implicated in familial rheumatoid arthritis.56 Although this finding does not prove that the same gene is responsible for both conditions, the findings are certainly suggestive.
One of the most interesting findings from our genome scan (Figure 4) has been the identification of a candidate locus for psoriasis on chromosome 16q.20 Although this locus remains to be confirmed, several groups presented evidence suggestive of a chromosome 16q locus at the May 2001 meeting of the International Psoriasis Genetics Committee in Stockholm, Sweden. Eight different genetic studies of Crohn disease, and 1 genetic study of AS have found linkage to the same region of chromosome 16.36 The risk for psoriasis in patients with Crohn disease is 7 times that of the general population.20 The NOD2 gene, located on chromosome 16q, has recently been implicated in Crohn disease.57 Preliminary studies from our laboratories have failed to implicate NOD2 in psoriasis.58 Nevertheless, this is an exciting development that may accelerate the identification of the 16q psoriasis locus. Another very recent study identified loci for childhood atopic dermatitis on chromosome regions 1q21, 3q21, 17q25, and 20p,59 all of which have been implicated in psoriasis by at least 1 study (Table 2). These chromosome regions are rather broad, and at this point we have no proof that the actual genes involved are the same in these various diseases. However, these findings raise the interesting possibility that certain genes may contribute to more than 1 disease, perhaps by controlling certain aspects of inflammation and immunity. They also illustrate how the many different ways that psoriasis can present in our patients (ie, guttate, pustular, inverse, with or without nail disease, arthritis, or inflammatory bowel disease) may reflect the particular mix of genes for psoriasis, arthritis, and/or gut disease that our patients are unfortunate enough to inherit.
FUTURE PERSPECTIVES
Much work remains to be done. Although there are intriguing clues to disease pathogenesis, eg, the mapping of PSORS1 in HLA, the Crohn disease connection, and the identification of the same candidate loci by multiple groups, none of the responsible genes have yet been identified. Our recent progress in the HLA region suggests that the identification of PSORS1 will occur soon. The characterization and function of PSORS1 and, it is hoped, the non-MHC psoriasis genes will unlock the mystery of psoriasis. These studies should provide appropriate molecular targets for improved drugs to treat the disease and key insights into how emotions, infections, and other environmental factors interact with genes to trigger this common, enigmatic disease.
AUTHOR INFORMATION
Accepted for publication June 20, 2001.
This research was supported by awards P30 HG00209-03 and R01 AR4274-01 from the National Institutes of Health, Bethesda, Md (Drs Elder, Nair, and Voorhees); by award DFG-WE 905/1-1 from the Deutsche Forschungsgemeinschaft, Bonn, Germany (Drs Henseler, Jenisch, and Christophers); by the Ann Arbor Veterans Affairs Hospital, Ann Arbor, Mich (Dr Elder); by the National Psoriasis Foundation, Portland, Ore, and by the Babcock Memorial Trust, Ann Arbor. Portions of these studies were conducted at the General Clinical Research Center at the University of Michigan, Ann Arbor, funded by grant M01EE00042 from the National Center for Research Resources, National Institutes of Health.
Our special thanks go to all the study subjects and referring physicians for their participation in these genetic studies of psoriasis during the past decade.
Corresponding author and reprints: James T. Elder, MD, PhD, 3312 CCGC, University of Michigan, 1500 E Medical Center Dr, Ann Arbor, MI 48109-0932 (e-mail: jelder{at}umich.edu).
From the Departments of Dermatology (Drs Elder, Nair, Chia, and Voorhees and Mr Stuart) and Radiation Oncology (Dr Elder), University of Michigan, and Department of Dermatology, Ann Arbor Veterans Affairs Hospital (Dr Elder), Ann Arbor, Mich; and the Departments of Dermatology (Drs Henseler and Christophers) and Immunology (Dr Jenisch), Christian-Albrechts University, Kiel, Germany.
REFERENCES
|
